Senior Full-Stack AI Engineer (RAG, Laravel, Next.js, Node, AWS)

We're building an AI-powered search and Q&A platform that lets teams query their own documents and internal data in plain language. We need a senior engineer who can own the AI layer and the application around it. To be clear about what this role is: the core of the work is production RAG. If your experience is limited to calling an LLM API through a wrapper library, this will be a frustrating fit. We need someone who has dealt with bad retrieval, argued about chunking strategy, measured whether a change actually improved answer quality, and watched a token bill get out of hand. You'll also be building the product around those AI features, so we need real full-stack depth alongside it. WHAT YOU'LL BE DOING Building RAG pipelines end to end: ingestion, chunking, embeddings, vector storage, hybrid retrieval, re-ranking, and evaluation Improving answer quality with measurement behind it, not guesswork Fine-tuning models where it's justified, and saying so when it isn't Building and maintaining backend services in Laravel and Node.js across a microservices setup Building front-end interfaces in Next.js and TypeScript Indexing and relevance tuning in Elasticsearch or OpenSearch Deploying and running everything on AWS with Docker and automated CI/CD Owning latency, cost, and reliability for the features you ship WHAT YOU NEED 8+ years building software professionally, with at least 3 years on LLM-based products Shipped RAG to production, including vector databases (pgvector, Pinecone, Qdrant, Weaviate, or OpenSearch k-NN) and LLM APIs (OpenAI, Anthropic, or AWS Bedrock) Laravel (PHP 8+) and Node.js, with solid API design and microservices experience Next.js, React, TypeScript Elasticsearch or OpenSearch: mappings, analyzers, relevance tuning AWS: ECS or EKS, Lambda, S3, RDS, SQS, IAM, CloudFront Docker and CI/CD pipelines (GitHub Actions, GitLab CI, or similar) MySQL or PostgreSQL, plus at least one NoSQL store Clear written English and the ability to explain trade-offs without hand-waving NICE TO HAVE Kubernetes and Terraform Python for ML and data tooling Agent orchestration (LangChain, LlamaIndex, or something you built yourself) LLM observability: tracing, token and cost monitoring Experience with document-heavy B2B SaaS products